Apache Camel CVE-2026-46591

Exploitation requires that a deployed Camel route using the camel-neo4j producer populates the CamelNeo4jMatchProperties map with attacker-controlled data - specifically that the attacker controls at least one JSON key name (not merely a value) in that map. The advisory identifies two realistic paths: a route that passes a request body directly as the match properties map, or a consumer that does not filter inbound Camel* headers and allows an external sender to supply the match map. The HTTP header-filter strategy applied to standard HTTP consumers explicitly blocks the CamelNeo4jMatchProperties header from being set by plain HTTP clients, so a vanilla REST endpoint is NOT vulnerable unless the route explicitly reads and forwards it. Exploitation is limited to the match/retrieve and deleteNode operations of the Neo4j producer; other Camel components are unaffected. No authentication level is determinable from available data without a CVSS vector, but access is constrained to whatever endpoint exposes the vulnerable route.

No CVSS vector or EPSS score is provided in the available intelligence, so risk must be assessed from the description and structural signals alone. The flaw is a full Cypher injection granting read, modify, and delete access to the entire Neo4j graph, which corresponds to a Critical-class impact in isolation. However, exploitation is gated by a significant architectural prerequisite: the vulnerable path is only reachable if an application route deliberately or inadvertently maps untrusted external data into the CamelNeo4jMatchProperties header. The HTTP header-filter strategy explicitly blocks this header from being set by plain HTTP clients, meaning it cannot be injected directly over a standard REST interface without a permissive or misconfigured route. This prerequisite meaningfully reduces the realistic attack surface compared to the raw severity of the impact. No KEV listing and no identified POC further support a lower urgency for most deployments. Priority should be highest for integrations that parse untrusted JSON bodies and forward keys into the match properties map - a pattern the advisory explicitly flags as vulnerable.

An attacker targets an API endpoint backed by a Camel route that passes a parsed JSON request body directly as the CamelNeo4jMatchProperties map. By sending a request body whose JSON keys contain Cypher syntax - such as a key named 'x] RETURN n UNION MATCH (m) DETACH DELETE m //' - the attacker causes Neo4jProducer to concatenate that key verbatim into the WHERE clause, altering the query structure and executing the injected Cypher. This results in deletion of all graph data or exfiltration of arbitrary nodes. No public proof-of-concept code has been identified, but the injection pattern is straightforward for an attacker familiar with Cypher syntax.